Many readers have asked for a review of the Scythe Infinity and we recently received the latest version of the famous Infinity heatpipe tower. The Infinity certainly qualifies as something of a legend, but legends have not always performed as expected in our cooling tests at AnandTech. Does the Infinity live up to the performance goalpost set by the Tuniq Tower 120, or like all the others tested so far, does it fall somewhat short? Our goal is to find an answer to this question.

Some early versions of the Infinity were praised for their cooling ability, but criticized for high noise by enthusiasts who value noise reduction above all else. This most recent version of the Infinity proclaims that it includes a "silent" 120mm fan. In our benchmarking, we will definitely examine the low-noise fan claim, but there is always the concern that a lower-noise fan may not cool as well as the highest CFM - and higher noise - fans. The Scythe fan is 120mm, which is large enough both to move a lot of air, and to do so at slow enough RPM to keep noise low.

Another addition to the newest Infinity is support for the AMD Socket AM2. Many kits like the Infinity, which were introduced before the launch of AM2, have not been updated to support the revised AM2 socket. Fortunately the Scythe Infinity is now one of the top-of-the-line coolers that fully supports AM2 - good news for AMD enthusiasts.

We have talked about universal mounting systems in the past, but Scythe comes through where many others have fallen short on that claim. Scythe includes adapters for any CPU you can currently buy.

Another feature of the Infinity is the ability to mount a 120mm fan on any of the 4 sides of the cooler. Some enthusiasts have found new ways to use this feature, by running two fan cross feed, two fan push-pull or even four fan configurations with dual push-pull. Frankly the Infinity is very large and heavy to start with and an Infinity with four fans would be daunting in both size and weight - and it actually would not fit in our case. However, the potential of a push-pull cooling arrangement intrigued us.

Scythe USA was also kind enough to send the latest Ninja Revision B, and the new revision of Ninja will be highlighted in an upcoming review. That gave us the advantage of a second Scythe cooler with the same silent fan. With two fans in hand we couldn't resist testing a dual-fan push-pull configuration, which would fit in the mid-tower case. The main interest was whether the two fans - one pushing air and one pulling air - would make any difference at all in the cooling abilities of the Infinity. The results are very interesting.

With this review of the Scythe Infinity, we are on the downslope of our tests of current high-end air coolers (Ed: with 50 plus inches of February snow where I live I couldn't resist the skiing metaphor). We have the Thermalright Ultra 120 in for review, and we will also be testing the Scythe Ninja Rev. B. Noctua, an Austrian manufacturer of cooling solutions, has also sent us the Noctua NH-U12F which will be reviewed in the near future. We will also take a closer look at the OCZ Vindicator, which is the recently introduced heatpipe tower from OCZ. Once these reviews are complete, plus any late additions to the air cooling Olympics, it will be time to select the best air cooler available.

Scythe Infinity


View All Comments

  • SmokeRngs - Wednesday, February 28, 2007 - link


    Sure it makes for interesting reading, so I guess I can't blame you too much, and the audience at Anandtech surely eats it up. I would just tell you I'd rather not have reviews cluttered up with this rather petty data.

    The information you wanted is there. Why are you complaining? There are plenty of people that don't care about the information you want but do want the information you deride. Both are there in a clear and concise manner for those that want to see one or both.
  • shabby - Monday, February 26, 2007 - link

    People who buy $50 heatsinks dont buy them because they're quiet, they buy them so they can overclock their cpu's. Plus they probably stick higher cfm fans, so they're not quiet anymore. Reply
  • Zambien - Monday, February 26, 2007 - link

    I disagree. When I bought my Zalman 7700cu for my current PC, the main reason was that it would provide similar cooling to my current HSF, with less noise. I didn't like the fact that my computer sounded like a jet engnine. I'm sure some people fall into the category, but others do not. Reply
  • cujo - Monday, February 26, 2007 - link

    why wouldn't you test with overclocks? Reply
  • arswihart - Monday, February 26, 2007 - link

    Not to mention they are talking about 50Mhz differences in OC to separate better and lesser performing heatsinks. Come on, that isn't even a significant difference. Did they repeat the result on multiple systems, or just the one. Does it matter? Of course not. Reply
  • arswihart - Monday, February 26, 2007 - link

    A heatsink doesn't do anything magic to give you higher OC's, it gives you lower temps, which will then let you OC higher.

    Measuring a CPU's OC is looking at data that is a step removed from what you should be looking at, which is simply the temperature. I could care less what they get to OC to, it will be almost certainly different for anyone who buys the heatsink.

    The only thing that I think is worth taking away from any heatsink review is: installation caveats, temps, and noise.
  • shabby - Monday, February 26, 2007 - link

    I find it funny that the infinity cant beat the tt120, does it have thicker heatpipes or what. What part of the tt120's construction make it better then other beefier/bigger coolers? Reply
  • Superdoopercooper - Monday, February 26, 2007 - link

    The first rule in A-B testing is to hold EVERYTHING constant except for the items being examined/tested.

    Therefore, in ALL of your heatsink reviews, the system down to the THERMAL COMPOUND should remain fixed, even if the heatsink ships with some "higher end" stuff.

    I think it would be wise to pick ONE thermal compound and use it on EVERY heatsink test. Then you are testing the performance of the heatsink (i.e. Heatsink #1 is better than Heatsink #2, with no exceptions), and not the thermal grease + heatsink.

    I would hate to think that heatsink #1 was the best, but only because it shipped with better compound than another. I think many people ponying up for these higher-$$ heatsinks will pony up $6 for some good thermal grease.

    Then, if you want to do an additional test that comments on the performance of the included thermal grease, that would/could be helpful to potential buyers.
  • Wesley Fink - Monday, February 26, 2007 - link

    I can easily link you to a review from a respected source that proves toothpaste and Kraft Vegemite are superior in cooling to Arctic Silver 5 :-) I have tested many thermal compounds and found little differnece among the quality products. That conclusion was a rude awakening for me. I have found much more variation in perfromance in how the thermal grease was applied than I have ever found in the thermal compounds themselves.

    Thus far, all of the tests have used our standard silver-colored (but no real silver content) tube thermal compound except the Thermalright MST 6775 and the Zalman pair. These came with top thermal grease, and yet none of the three beat our Tuniq or this Infinity. If the cooler company cares enought to supply a premium thermal compound we test with that compound.
  • Superdoopercooper - Monday, February 26, 2007 - link

    If it is true that all thermal compounds are nearly equal, then I guess that's fine. I would like to see the link, actually.

    That still doesn't change the fact that the best A-B testing holds EVERYTHING constant except the single item being tested.

    Since AT is the site I hold in the highest regard in terms of info on computer components, I thought I'd just throw up my $0.02 on ways to maybe make these tests "better" and/or more scientific.

    Could just be my test engineering background talking. :-P

Log in

Don't have an account? Sign up now